Phase dependent differential thermopower of SND junctions: Pair-breaking effects and Gaussian fluctuations

to appear in Phys. Rev. B

Phase dependent differential thermopower of SND junctions: Pair-breaking effects

and Gaussian fluctuations

0^

Sergei Sergeenkovd and Marcel Ausloos
SUPRAS, Institute of Physics, University of Liege, B-4000, Liege,

(February 1, 2008)

oo

o

Oh'

X3

o

>

in

(N
(N

O

a^

OS
-I— >

I

O

o

X

We start with revisiting our previous results on ther-
moelectric response of SNS configuration in a C-shaped
BixPhi-xSr2CaCu20y sample in order to include strong
fluctuation effects. Then, by appropriate generalization of
the Ginzburg-Landau theory based on admixture of s-wave
(S) and d-wave (D) superconductors, we consider a differential
thermoelectric power (TEP) of SND junction. In addition to
its strong dependence on the relative phase 9 = (f)s — 4)4 be-
tween the two superconductors, two major effects are shown
to influence the behavior of the predicted TEP. One, based
on the chemical imbalance at SD interface, results in a pro-
nounced maximum of the TEP peak near 6 — -k 12 (where the
so-called s -f id mixed pairing state is formed) for two iden-
tical superconductors with Ted = T^s = Tc. Another effect,
which should manifest itself at SD interface comprising an
s-wave low- Tc superconductor and a d-wave high- Tc super-
conductor with Ted ^ Tea, predicts Sp oc Ted — Tea for the
TEP peak value. The experimental conditions under which
the predicted behavior of the induced differential TEP can be
measured are discussed.

PACS numbers: 74.25.Fy, 74.80.Fp

I. INTRODUCTION

During the last few years the order parameter sym-
metry has been one of the intensively debated issues in
the field of high- Tc superconductivity (HTS). A number
of experiments points to its (ij,2_j^2-wave character [nl.
Such an unconventional symmetry of the order param-
eter has also important implications for the Josephson
physics because for a d-wave (D) superconductor the
Josephson coupling is subject to an additional phase de-
pendence caused by the internal phase structure of the
wave function. The phase properties of the Josephson
effect have been discussed within the framework of the
generalized Ginzburg-Landau (GL) 0] as well as the tun-
neling Hamiltonian approach [g[. It was found Q that
the current-phase relationship depends on the mutual ori-
entation of the two coupled superconductors and their in-
terface. This property is the basis of all the phase sensi-
tive experiments probing the order parameter symmetry.
In particular, it is possible to create multiply connected
d-wave superconductors which generate half-integer flux
quanta as observed in experiments ^. Various interest-
ing phenomena occur in interfaces of d-wave supercon-

ductors. For example, for an interface to a normal metal
a bound state appears at zero energy giving rise to a zero-
bias anomaly in the /-^-characteristics of quasiparticle
tunneling WM while in such an interface to an s-wave
(S) superconductor the energy minimum corresponds to
a Josephson phase different from or tt. By symmetry, a
small s-wave component always coexists with a predom-
inantly d-wave order parameter in an orthorhombic su-
perconductor such as YBCO, and changes its sign across
a twin boundary pjy]. Besides, the s-wave and d-wave
order parameters can form a complex combination, the
so-called s ± id-state which is characterized by a local
breakdown of time reversal symmetry T either near sur-
faces M-Ol or near the twin boundaries represente d by
tetragonal regions with a reduced chemical potential [Q .
Both scenarios lead to a phase difference of ±7r/2, which
corresponds to two degenerate states ||lj,|l^. Moreover,
the relative phase oscillations between two condensates
with different order parameter symmetries could mani-
fest themselves through the specific collective excitations
("phasons") ^.

At the same time, a rather sensitive differential tech-
nique to probe sample inhomogeneity for temperatures
just below Tc, where phase slippage events play an im-
portant role in transport characteristics has been pro-
posed |1^ and successfully applied Q for detecting
small changes in thermoelectric power (TEP) of a spec-
imen due to the deliberate insertion of a macroscopic
SNS junction made of a normal-metal layer N, used to
force pair breaking of the superconducting component
when it flows down the temperature gradient. Analy-
sis of the thermoelectric effects provides reasonable es-
timates for such important physical parameters as the
Fermi energy, Debye temperature, interlayer spacing etc.
In particular, a carrier-type-dependent thermoelectric
response of such a SNS configuration in a C-shaped
BixPbi^xSr2CaCu20y sample has been registered and
its temperature behavior below Tc has been explained
within the framework of GL theory ||l^ .

In the present paper, we consider theoretically the
behavior of induced TEP at NS, ND, and SD inter-
faces and discuss its possible implications for the above-
mentioned type of experiments. The paper is organized
as follows. In Section II we briefly review the experi-
mental results for SNS configuration (with both hole-
like and electronlike carriers of the normal-metal N in-
sert) and present a theoretical interpretation of these re-

suits, based on GL free energy functional, both below
and above T^- The crucial role of the pair-breaking ef-
fects (described via the chemical balance A/j, between
the quasiparticles and Cooper pairs) in understanding
the observed phenomena is emphasized. In Section III,
extending the early suggested [|ll|Jl4| ] GL theory of an
admixture of s-wave and d-wave superconductors to in-
corporate strong pair-breaking effects, we calculate the
differential thermopower AS' of SND configuration near
Tc- The main theoretical result of this Section is the
prediction of a rather specific dependence of AS* on rel-
ative phase shift 9 — (j)s — (j^d between the two super-
conductors. Two independent mechanisms contributing
to the peak value Sp{6) = AS{Tc,9) of the differential
thermopower are discussed. One, based on the chemi-
cal balance between S and D superconductors at an SD
interface (and responsible for charge-related interference
effect), is discussed in Section IIIA. It results in a pro-
nounced maximum of the peak Sp{9) near 9 = 7r/2 (the
so-called s + id mixed pairing state) for two identical su-
perconductors with Ted = Tcs = Tc- This mechanism
can be realized, e.g., in a d-wave orthorhombic sam-
ple (like YBCO) with twin boundaries which are rep-
resented by tetragonal regions of variable width, with
a reduced chemical potential. Another mechanism (dis-
cussed in Section IIIB), which is active in the absence of
the normal- metal layer, takes place when two different
superconductors with T^d ^ Tcs are used to form an SD
interface. This situation can be realized for an s-wave
low-Tc superconductor (like Pb) and a d-wave high-T^
superconductor (like orthorhombic YBCO) and is shown
to yield Sp{9) oc Ted — Tcs for predicted TEP peak value.

II. SNS CONFIGURATION REVISITED

A. Experimental setup and main results

Before turning to the main subject of the present pa-
per, let us briefly review the previous results concern-
ing a thermoelectric response of SNS configuration in a
C-shaped BixPhi-xSr2CaCu20y sample (see Ref.l9 for
details). The sample geometry used is sketched in Fig.l,
where the contact arrangement and the position of the
sample with respect to the temperature gradient V^T
is shown as well. Two cuts are inserted at 90° to each
other into a ring-shaped superconducting sample. The
first cut lies parallel to the applied temperature gradient
serving to define a vertical symmetry axis. The second
cut lies in the middle of the right wing, normal to the
symmetry axis, separating two similar superconductors
with 5" = S" = S or D and completely interrupting the
passage of supercurrents in this wing. The passage of any
normal component of current density is made possible by
filling up the cut with a normal metal N. The carrier
type of the normal-metal insert A'' was chosen to be ci-
ther an electronlike Ne (silver) or holelike N^ (indium).

-S'NS"

'2 T

7 6 5

FIG. 1. Schematic view of the sample geometry with
S'A5'"-junction and contacts configuration. Here S' and S"
stand for s-wave and/or d-wave type superconductor. The
thermopowers Sr and Sl result from the thermal voltages
detected by the contact pairs 4 — 5 and 1 — 7, respectively.

The measured difference between the thermopowers of
the two halves AS* = Sr — Sl was found to approximately
follow the linear dependence

AS{T)^Sp±B^{Tc~T),

(1)

with slopes B' and B+ defined for T < T^ and T > T^,
respectively. Here Sp = AS'(Tc) is the peak value of
AS{T) at T = Tc. The best fit of the experimental
data with the above equation yields the following val-
ues for silver (Ag) and indium (In) inserts, respectively
(see Fig.2): (i) Sp{Ag) = -0.26 ± 0.01/Lty/A', B-{Ag) =
-0.16 ± O.l^y/if^ B-{Ag)/B+{Ag) = 1.9 ± 0.1; (ii)
Sp{In) =0.83±0.01/iV"/A:, B-(/n) = 0.17± 0.1^^^/^:^,
B- [In) I B+ [In) = 2.1 ± 0.1.

B. Interpretation

It is important to mention that, unlike the case of
mixed SND configuration (considered in Section III), the
suggested interpretation of the current experimental re-
sults for SNS configuration does not involve the phase of
the order parameter and hence is not sensitive (at least
near Tc) to the pairing symmetry of the two supercon-
ductors S" and S" . To describe the observed behavior
of the differential TEP both below and above T^ we can
roughly present it in a two-term contribution form [n9 20

AS{T)^ASa.{T) + ASfi{T),

(2)

where the average term ASav{T) is assumed to be non-
zero only below Tc (since in the normal state the TEP of
HTS is found to be very small ||2l|,||l) while the fluctu-
ation term A5//(T) should contribute to the observable

AS{T) for T ~ Tc- In what follows, we shall discuss these
two contributions separately within a mean-field theory
approximat ion .

90 100 110 120 130 140
Temperature (K)

FIG. 2. The temperature dependence of the observed
differential thermopower of SNS configuration defined by
Eq.(l). The upper (lower) part of the picture refers to In
(Ag) normal-metal insert in the right wing (see Fig.l). The
asymmetric curved triangle shapes are approximated by lin-
ear shapes produced by the linear fit to the data points (see
the text for details).

1. Mean value of the differential thermopower: ASav{T)

Assuming that the net result of the normal-metal in-
sert is to break up Cooper pairs that flow toward the
hotter end of the sample and to produce holelike (In) or
electronlike (Ag) quasiparticles, we can write the differ-
ence in the generalized GL free energy functional AQ of
the right and left halves of the C-shaped sample as

Ag[i'] = ATiij] + A/xlV-l^

where

and

AT[^j] =Tr~Tl= a(r)|^p + ^IV

A/i = [iB. - [i-L-

(3)

(4)

(5)

Here ^ = li/jle*"^ is the superconducting order parameter,
A[i accounts for the chemical balance between quasipar-
ticles and Cooper pairs; a{T) = a{T — Tc) and the GL
parameters a and f3 are related to the critical temper-
ature Tc, zero-temperature BCS gap Aq = l.TG/csTc,
the in-plane Fermi energy E'j^ = p'p/m'^ij, and the
total particle number density n as a = /3n/2Tc =
{AAokB / E'^){me/m'^ij) . In fact, in layered supercon-

ductors, Afi ~ E'p c^ Jc/Ep, where E$,

Ef

I In

the out-of-plane Fermi energy and Jc the interlayer cou-
pling energy within the Lawrence-Doniach model {"fm =
m*/m'^i^ is the mass anisotropy ratio, and tti*^ ~ Srric for
this material).

As usual, the equilibrium state of such a system is de-
termined from the minimum energy condition dQ/d\iJj\ =
which yields for T < Tc

IV'o

ajTc -T)-Afi

(6)

Substituting |'0oP into Eq.(3) we obtain for the average
free energy density

AniT) . AGi^Po] = _ K^--J)-^A^]^ (7)

2/3

In turn, the difference of thermopowers
be related to the corresponding difference
entropies Act = -dAVt/dT as AS{T) =
where q is the charge of the quasiparticles.
the mean value of thermopower associated
breaking event reads (below Tc)

AS{T) can
of transport
Aa{T)/nq,
Thus finally
with a pair-

Ajav\-L ) — i>p,av J^av[-'^c ^ ji

with

and

ijp^av

A^
2^

B„

ApkB
2qEfTc

(8)

(9)

(10)

Before we proceed to compare the above theoretical find-
ings with the available experimental data (see Fig. 2), we
first have to estimate the corresponding fluctuation con-
tributions to the observable TEP difference, both above
and below Tr.

2. Mean-field Gaussian fluctuations of the differential
thermopower: ASfi{T)

The inffuence of superconducting fluctuations on trans-
port properties of HTS (including TEP and electri-
cal conductivity) has been extensively studied for the
past few years (see, e.g., ||2^-|3^ and further references

therein). In particular, it was found that the fluctuation-
induced behavior may extend to temperatures more than
lOK higher than the respective Tc. Let us consider now
the region near Tc and discuss the Gaussian fluctuations
of the pair-breaking-induced differential TEP ASfi{T).
Recall that according to the theory of Gaussian fluctu-
ations 1^^, the fluctuations of any observable, which is
conjugated to the order parameter -0 (such as heat capac-
ity, susceptibility, etc) can be presented in terms of the
statistical average of the square of the fluctuation ampli-
tude < ((5'0)^ > with Sip = ip — ipQ. Then the differential
TEP above {+) and below (— ) Tc have the form of

AS%{T)^A<iSi;f>±,

where

< (Si^)^ >=

^ / d|V'|(<5V')2e-^[^l

(11)

(12)

Here Z = J d\^\e ^^^'^ is the partition function with
i;[V'] = (Agfij] - Ag[V'o])/fci3T. a is a coefficient to be
defined below. Expanding the free energy density func-
tional At/['0]

Ag[ij]

Ae[^o] + I

d^AG

dij^

{H?

(13)

IV'hI'Aol

around the mean value of the order parameter -00 , which
is defined as a stable solution of equation dQ /d\4!\ —
we can explicitly calculate the Gaussian integrals. Due
to the fact that |V'oP is given by Eq.(6) below Tc and
vanishes at T > Tc, we obtain finally

AS^iiT)

AkeTc

4a(Tc-T)-4A/i'

and

A5+(T)

AkgTc

2a{T - Tc) + 2An'

T <Tr

T>Tc

(14)

(15)

Then the above equations bring about the following ex-
plicit expression for the constant parameter A, namely
A = {AA^/'ikBTc)Sp.av This in turn leads to the fol-
lowing expressions for the fluctuation and total contribu-
tions to peaks and slopes through their average counter-
parts (see Eqs.(9) and (10)): 5*+^; = S'p = (2/3)5p,a^,

Sp fi ~ — (l/3)5p,au, B^i = (l/3)i?ai., -Bj; = (2/3)i3a'u,

B- = Bav + BJi = iA/3)Bav, and B+ = B+ = (2/3)Ba..
Thus, in agreement with the observations, B~ /B^ = 2
independent of the carrier type of the normal-metal in-
sert. Let us proceed to discuss separately the case of In
and Ag inserts.

a. N = In (holelike metal insert). In this case, the
principal carriers are holes, therefore q = +e. Let the
holclike quasiparticle chemical potential (measured rel-
ative to the Fermi level of the free-hole gas) be posi-
tive, then fiq = +11 and Afi = fi + 2fi = 3fi (two holes
come from condensate and one hole is brought by normal
metal). Therefore, for this case Eq.(l) takes the form

AS'\T) = Spiln) ± B^{In){Tc - T),

where

Sp{In)

/^

ks
e J \ ksTc

(19)

(20)

and

B-{In) ^

2AokB_
3eEfTc '

B+iIn)^-B-iIn). (21)

b. N = Ag (electronlike metal insert). The principal
carriers in this case are electrons, therefore q = —e. The
electronlike quasiparticle chemical potential jiq = — /i.
Then A^ = —^ + 2fi — /i (plus one electron means minus
one hole). For this case Eq.(l) takes the form

AS%T) = Sp{Ag) ± B^{Ag){Tc - T), (22)

where

As we shall see below, for the experimental range of
parameters under discussion, A^{Ep / Aq) ^ kB\Tc —
T\. Hence, with a good accuracy we can approximate
Eqs.(14) and (15) as follows

asUt)

s^

pj'l

±B%{Tc

T),

where

S,

pji

AksTc
4A/X

B

fi

AkgTca
T2A^'

and

S.

pji

-25,

pj"

B

fi

2B

fl-

ue)

(17)

(18)

Furthermore, it is quite reasonable to assume that S =

Op — Dp, wnerc j^ — ^p^av

^pji ^^'^ ^p — ^pji-

Sp(

Ag) = -

and

B-

(Ag)

= -

2AokB
SeEfTc

M

ikBTcJ '

(23)

B+{Ag) = -B-{Ag). (24)

By comparing the obtained theoretical expressions with
the above-mentioned experimental findings for the slopes
B^ and the peak Sp values for the two normal-metal in-
serts (see Fig. 2), we can estimate the order of magnitude
of the in-plane Fermi energy E'j}' and interlayer coupling
energy Jc. The result is: E'j}' ~ 0.16ey and Jc — 4:meV,
in reasonable agreement with the other known estimates
of these parameters pO| . In turn, using these parame-
ters (along with the critical temperature), we find that
Jc/ksAo ~ lOOK. This justifies the use of the linearized

Eq.(16) for the temperature interval \Tc - T\ ■€. IQQK.
As is seen in Fig. 2, the observed differential TEP practi-
cally disappears already for \Tc — T\ > IQK. Moreover,
as it follows from Eqs.(20) and (23), the calculated ratio
for peaks \Sp{In)/ Sp{Ag)\ = 3 is very close to the cor-
responding experimental value \S'jf^{In)/S'jf^{Ag)\ =
3.2 ± 0.2 observed by Gridin et al |1^. Finally, as it
follows from the above analysis, the calculated slopes
B~ below Tc for the two metal inserts coincide with
each other, namely B~{In) — —B~{Ag), and are twice
their counterparts above Tc, i.e., B~{In) — 2B^{In) and
B~{Ag) = 2B~^{Ag), in a good agreement with the ob-
servations. It is worthwhile to note that a very similar
behavior of the induced TEP (including peaks and slopes
both above and below Tc) has been observed in strong ap-
plied magnetic fields EG]. In fact, replacing the chemical
potentials difference A/i (responsible for pair-breaking ef-
fects in SNS junction) in the above equations by (IbH
term (where fiB is the Bohr magneton and H the applied
magnetic field) we recover most of the formulas presented
in Ref.20 where magneto-TEP of Bi2Sr2CaCu20y su-
perconductors was studied.

III. SND CONFIGURATION: PREDICTION

Since Eqs.(3)-(5) do not depend on the phase of the
order parameter, they will preserve their form for a
DND junction (created by two d-wave superconductors,
S' — S" = D, see Fig.l) as well, bringing about the
results similar to that given by Eqs.(8)-(10). It means
that the experimental method under discussion (and its
interpretation) can not be used to tell SNS and DND
configurations apart, at least for temperatures close to
Tc- As for low enough temperatures, the situation may
change drastically due to a markedly different behavior
of s-wave and d-wave order parameters at T <^ Tc (where
the node structure begins to play an important role). As
we shall show below, this method, however, is quite sen-
sitive to the mixed SND configuration (when S' = S has
an s-wave symmetry while S" — D is oi a, d-wave symme-
try type, see Fig.l) predicting a rather specific relative
phase {9 = 4>s— (j)d) dependencies of both the slope B{d)
and peak Sp{9) of the observable thermopower difference
AS{T,

and

Following Feder et al |14|, who incorporated chemical
potential effects near twin boundaries into the approach
suggested by Sigrist et al ||ll| , we can represent the gen-
eralized GL free energy functional AQ for SND configu-
ration of the C-shaped sample in the following form

Ag[i:s,H = Ag[^,] + Agi^jM] + Ag„t, (25)

where

A(?[v^,] = ^(^)|v,p + %|v-,|^

Ag[M^Aa{T)\M^ + ^\i;d\\

(26)

(27)

(28)

Here ipn — IV'nle"^" is the n-wave order parameter (n ~
{s,d}); An(T) = a„(r) -|- A/x„ where a„(r) = a„(T -
Ten) with the corresponding parameters a„, /3„, Ten, and
Afin for s-wave and d-wave symmetries.

An equilibrium state of such a mixed system is deter-
mined from the minimum energy conditions ^g/^\^l>s\ =
and dg/dlipdl = which result in the following system
of equations for the two equilibrium order parameters ipso
and ipdo

As\iPso\ + /3s\^so\^ + r{e)\i,so\\i^do\^ --

Ad\4'do\ + M^dol^ + n0)\iPdo\\Ao\''

where we introduced relative phase 6 —
parameters

A{0)\i;do\
= A(0)|V^o|

(29)
(30)

T{e).

A{0)

7i + 72 cos 29
- 6i + S2 cos 9

bd dependent
(31)

Notice that unlike chemical potentials difference A/x„
(which is responsible for pair-breaking effects in SND
junction due to the normal-metal insert), the interference
terms (5i_2 describe the chemical balance between s-wave
and d-wave superconductors at SD interface in the ab-
sence of a normal-metal layer. Therefore, the effects due
to A/i„ ^ should be distinguished from the interference
effects due to A{9) ^ 0. The latters are generically close
to the interference effects between the two condensates
and are described by the r(6') term. Notice also that the
A{9) term favors 9 — lir {I integer), while the T{9) term
favors 9 = wk /2 {n — 1,3,5...) which corresponds to a
T- violating phase ||lj]. In principle, we can resolve the
above system (given by Eqs.(29)-(31)) and find ipno for
arbitrary set of parameters a„ , /3„ , and Ten ■ For simplic-
ity, in what follows we restrict our consideration to the
two limiting cases which are of the most importance for
potential applications.

A. Twin boundaries in orthorhombic d-wave
superconductors

1. Mean value of the differential thermopower: ASav{T,8)

First, let us consider the case of similar superconduc-
tors comprising the SND junction with [ipsol = iV'dol =
IV^ol, as^ad = a, Ps = Pd = P, A/x^ = A^^ = A/i, and
Tcs = Ted = Tc- This situation is realized, for example,
in a d-wave orthorhombic sample (like YBCO) with twin
boundaries which are represented by tetragonal regions
of variable width, with a reduced chemical potential WM .
In this particular case, Eqs.(29) and (30) yield for T < Tc

l^oP -

(3 + T{0)

(32)

where A/i(0) = A/x — A{6). After substituting the thus
found IV'o I into Eq.(25) we obtain for the gcnerahzcd equi-
hbriuni free energy density

An{T,e) = Ag[i^o]^

(33)

which in turn results in the following expression for the
mean-field value of the thermopower difference in a C-
shaped sample with SND junction (see Fig.l)

ASav{T, 9) = Sp,avi9) - BaMiTc - T),

where

and

^p,av\P) —

qTc

Bav{d)

qTc

A/i(g)

(3 + Tie)

(3

P + T{d)

(34)

(35)

(36)

2. Mean-field Gaussian fluctuations of the differential
thermopower: ASfi (T, 6)

Following the lines of Section II, we can present the
fluctuation contribution to the differential TEP above
(+) and below (— ) Tc as

ASfiiT,9) = A{9)[< (#,)2 >± + < {di^df >± (37)

+2 < S^psStpd >±],

where, e.g..

>-

z

d\^,\ / d\M{5^sfe-''^^-^-\ (38)

Here Z — j d\ipg\ j' d\ipd\G~^^'^''^'^'^ is the correspond-
ing partition function with S[^s,^c;] = {AQ[iIj g^ipd\ —
A5[i/'s07 'fpdo]}/kBT. A{9) is a coefficient to be fixed later.
Expanding the free energy density functional AQlipg, ipd]

Agi^syipd] ~ Ag[^so,iPdo]

'd^Ag'

(39)

d^Ag

{Hs?

\ips\ = \^sa\

drd

d^Ag

dipsdtpd

(S^d)

\-4'd\=\i>do\

(StpsS-ipd),

\i'^\=\4'^o\

around the mean values of the order parameters tpnOi
defined as stable solutions of equations dg/d\ipn\ = we
can explicitly calculate the Gaussian integrals to obtain

< {S^Psf > =< {Hdf >-

(40)

keTcP

4{(3-r)[aiTc-T)-Af,{9)y

< SipsS^pd >-= -

kBTcT

AiP-T)[a{Tc-T)-A^^i9)]

and

< {Si^sf >+ =< (S^Pdf >+

knTr

2[a{T-Tc)+A^i{9)y

< dipsStl^d >+= 0,

(41)

(42)

(43)

for the order parameters fluctuations below and above
Tc, respectively. In principle, the above expressions
completely determine the fluctuation contribution to the
seeking TEP of SND contact in the presence of strong
iV-metal induced pair-breaking effects. However, to com-
pare it with the earlier calculated mean-field values, let
us assume that [An{e)]{EF/Ao) > feslTc - T\. Then,
with a good accuracy we can approximate Eqs.(40)-(42)
as follows

ASfi (T, 6) c, 5±^, (9) ± B% (9) (T, - T) , (44)

where

SpJi((^) = -

AkeTc
2An{9)'

BniO)

AksTcOi

mm?

(45)

(46)

and

S;ji{9) = -2S;ji{9), B+i9)^2BJ^i9). (47)

Again, requiring that Sp{9) = Sp{9) = Sp{9), where

Sp -

give

'^p — Sp,av + Spji and 5+ — S^j^, the above equations

A{9)

SqkBT^ [/? + Ti9)]

(48)

for the above- introduced parameter (see Eq.(37)). This
in turn leads to the following expressions for the fluctua-
tion and total contributions to peaks and slopes through
their average counterparts (Cf. Section II): St' fi =

Op ^ [Zf ojOp,av^ O j;^ = [1/ o)Op,av: ^ fl ^ \^l ^)t>av:

B+ = (2/3)Ba., B- - Bav + BJi - (4/3)Sa„, and
B+ = B+ = {2/3)Bav Thus, the ratio B-{9)/B+{9) =
2 remains universal showing no dependence on the rel-
ative phase difference 9. As expected, completely ne-
glecting the interference terms (when both r(6') ^- and

A{9) -^ 0) we recover all the results of Section II for
two independent order parameters. Finally, the differen-
tial TEP of SND junction consisted of two superconduc-
tors with similar critical parameters but markedly differ-
ent pairing symmetries (like in a d-wave orthorhombic
YBCO with s-wave tetragonal twin boundaries) reads

AS{T,9) - Spi0) ± B^i0)iT, ~ T),

with

SpiO)

2(352

6 + cos (

3gTc72 \ 7 -h cos 26*

and

B-{9) =2B+{9) =

Aa

(3

'?>qTcl2 V7 + C0S26'
Here 7 = (/3 -f 7i)/72 and 6 = (A^ + 5i)/52-

(49)

(50)

(51)

FIG. 3. Predicted phase-dependent thermopower response
of SND configuration in a C-shaped sample (see Fig.l). Solid
and dashed lines depict, respectively, the relative phase 6
dependence of the normalized slope B~ (6) / B~ (Q) and peak
value Sp{d)/Sp{0) of the induced thermopower difference, ac-
cording to Eqs.(50) and (51).

Fig. 3 shows the predicted 6'-dependent behavior of the
normalized slope B~{9)/B~{0) (solid line) and the peak
Sp{9)/Sp{0) (dashed hue) of the fi-iVD-induced ther-
mopower difference AS{T,9) just below Tc, for •y = S =

2. As is seen, both the slope and the peak exhibit a max-
imum for the s + id state (at 9 = 7r/2) and a minimum
for the s — d state (at 9 ^ tt). Such sharp dependen-
cies suggest quite an optimistic possibility to observe the
above-predicted behavior of the induced thermopower,
using the sample geometry and experimental technique
described in Section II. Besides, when the pair-breaking
effects (due to the normal-metal insert in SND junc-
tion) are negligible (so that A/i = 0), Eqs.(49)-(51) will
describe the differential TEP at the SD interface where
the pair-breaking interference effects (governed by the
A{9) term) will dominate its peak behavior. This situa-
tion would allow one to get a more detailed information
about the mixed pairing states and the introduced phe-
nomenological parameters 71.2 and 61,2-

B. Low- Tc s-w^ave superconductor and high- Tc d-wave
superconductor

Let us turn now to another limiting case and consider
an SD interface formed by two different superconductors
(with iV'sol ¥" iV'dol, as ^ ad, f3s ^ Pd, and T^s ^ Ted) in
the absence of a normal-metal layer (which is responsible
for pair-breaking effects). We shall also assume that the
charge-related interference effects (governed by the A{9)
term) are rather small and can be safely neglected. Thus,
in this Section we consider the situation when A/i„ =
and A(9) = 0. Such a situation can be realized for an s-
wave low- Tc superconductor (like Pb) and a d-wave high-
Tc superconductor (like orthorhombic YBCO) [yH]. In
fact, the solution for this particular case is well-known.
It has been discussed by Sigrist et al ||ll| in a somewhat
different context. In principle, we can obtain both an av-
erage and fluctuation contributions to the resulting TEP
for this case, following the recipes of the previous Section.
And in particular, it can be shown that the fluctuation
contribution is still governed by expressions similar to
the ones given by Eqs.(40)-(43) with an evident change
in parameters, a — > a„ and (3 —^ (3n for s- and d-wave su-
perconductors. Since, however, the correponding expres-
sions are rather cumbersome, in what follows we restrict
our analysis with the average values of the induced TEP
only.

Assuming Tcs < Ted, two temperature regions should
be distinguished.

a. T < T*{9). In this region, the corresponding ex-
pressions for the equilibrium order parameters read (see
Eqs.(29) and (30))

1^,

sOl

,2 (3dasiT)-r{9)ad{T)

T^9)-Ps(3d

and

^^_ f3sad{T)-T{9)as{T)
''^'°' " r^9)-(3s(3d '

(52)

(53)

where the transition point T* [6) , defined by the equation
ipso{T*) = 0, is strongly ^-dependent and deviates from
an s-wave critical temperature Tcs as follows

T*{e)^T,,

asPd - adT{0) '

(54)

where ATc = Ted - Tcs-

After substituting the solution given by Eqs.(52) and
(53) into Eq.(25) we obtain for the average thermopower
difference

:ATc)-BUe)[T*iO)-T],

(55)

ASUt,0;ATc) = SI,

where

SpnviO'^^T,

2qN

o^l^Tc

and

BiM -

OisPd ~Oid^{d)_

2asadT{e) - a^Pd - a^Ps

(56)

(57)

Here N — UsUd/ {n
density.

h. T*{e) <T <Tcd
Eqs.(29) and (30)

2qN[T{e)^ - PsPd] ■

Ud) is a generalized carrier number

In this region we obtain from

IfAsol = 0, \iJdo\

adjTcd - T)
/3d

(58)

for the equilibrium order parameters. And the resulting
mean-field thermopower difference in this region is

^Sii{T,0)

S" (

ATc) + Bii[T-T*{9)l (59)

the predicted behavior of the induced thermopower at
such SD interface using a low- Tc s-wave and a high- Tc
d-wave superconductors (like, e.g., Pb and YBCO).

„0.8

tD

T7T

cd cs

FIG. 4. The ratio T*{9)/Tcs as a function of T^djTcs for
different Q calculated according to Eq.(54). Solid, dashed,
and dotted lines correspond to Q = -n i^s — d state), Q — 7r/2
(s + id state), and 9 = n/4, respectively.

where

Sp.aviO'^^Tc

2SL

\ATc

B

II

qNpd

(60)

Figure 4 depicts the ratio T*{6)/Tcs as a function of
Tcd/Tcs for different 0. As we can see, for the chosen
set of parameters (71 = 72 = /?« and Ud — n,,), in the
mixed s + id pairing state (with 9 = 7r/2, dashed line)
T*(7r/2) = Tcs for aU Tcd/Tcs- As it follows from Eqs.(56)
and (57), this state is described by the following depen-
dencies of the TEP peak and slope (below T*)

S"^ C

'-^p^av \

lATc)

2Arfofc£

1

T

J- cs

Ted

and

Bi

2'

1

E^

(61)

(62)

As it is evident from the above equations, in this regime
the peak's amplitude Sp is entirely dominated by the crit-
ical temperatures difference Ted — Tcs of the two super-
conductors while the slope B is governed by the corre-
sponding Fermi energies. It would be interesting to test

IV. CONCLUSION

In summary, to probe into the pairing state of high-
Tc superconductors, we calculated the differential ther-
mopower AS of SND junction in the presence of strong
pair-breaking effects (due to the normal-metal layer N)
and charge-related interference effects (due to the chem-
ical imbalance at SD interface) using the generalized
Ginzburg-Landau theory for an admixture of s-wave and
rf-wave superconductors near Tc- The calculated ther-
mopower was found to strongly depend on the relative
phase 6 ~ (ps ^ (pd between the two superconductors (ex-
hibiting a pronounced maximum near the mixed s + id
state with 9 — 7r/2) and their critical temperatures. The
experimental conditions under which the predicted be-
havior of the induced thermopower could be observed
were discussed.

ACKNOWLED GMENTS

We thank J. Annctt, J. Clayhold and T.M. Rice for
their interest in this work and very useful discussions.
S.S. was financially supported by FNRS (Brussels, Bel-
gium). M.A. was financially supported by the Minister of
Education under contract No. ARC (94-99/174) of ULg.

* e-mail: serge@gw.unipc.ulg.ac.be
[1] D.J. van Harlingen, Rev. Mod. Phys. 67, 515 (1995).
[2] M. Sigrist and T.M. Rice, Rev. Mod. Phys. 67, 503

(1995).
[3] C. Bruder, A. van Otterlo, and G.T. Zimanyi, Phys. Rev.

B 51, 12904 (1995).
[4] S. Yip, Phys. Rev. B 52, 3087 (1995).
[5] C.C. Tsuei, J.R. Kirtley M. Rupp, J.Z. Sun, A. Gupta,

M.B. Ketchen, C.A. Wang, Z.F. Ren, J.H. Wang, and M.

Blushan Science 271, 329 (1996).
[6] C.-R. Hu, Phys. Rev. Lett. 72, 1526 (1994).
[7] Y. Tanaka and S. Kashiwaya, Phys. Rev. B 53, R11957

(1996).
[8] M.B. Walker, Phys. Rev. B 53, 5835 (1996).
[9] K.A. Kouznetsov, A.G. Sun, B. Chen, A.S. Katz, S.R.

Bahcall, John Clarke, R.C. Dynes, D.A. Gajewski, S.H.

Han, M.B. Maple, J. Giapintzakis, J.-T. Kim, and D.M.

Ginsberg, Phys. Rev. Lett. 79, 3050 (1997).
[10] M. Sigrist, D.B. Bailey, and R.B. Laughlin, Phys. Rev.

Lett. 74, 3249 (1995).
[11] M. Sigrist, K. Kuboki, P.A. Leo, A.J. MiUis, and T.M.

Rice, Phys. Rev. B 53, 2835 (1996).
[12] K. Kuboki and M. Sigrist, J. Phys. Soc. Jpn. 65, 361

(1996).
[13] A. Huck, A. van Otterlo, and M. Sigrist, Phys. Rev. B

56, 14163 (1997).
[14] D.L. Feder, A. BeardsaU, A.J. Berlinsky, and C. Kallin,

Phys. Rev. B 56, R5751 (1997).
[15] A.B. Kuklov, Phys. Rev. B 52, R7002 (1995).
[16] M. Matsumoto and H. Shiba, J. Phys. Soc. Jpn. 64, 3384

(1995); ibid. 64, 4867 (1995).
[17] P.V. Shevchenko and O.P. Sushkov, Phys. Lett. A236,

137 (1997).
[18] V. Gridin and W. Datars, Phys. Rev. B 43, 3675 (1991).
[19] V. Gridin, S. Sergeenkov, R. Doyle, P. de Villiers, and

M. Ausloos, Phys. Rev. B 47, 14594 (1993).
[20] S. Sergeenkov, V. Gridin, P. de Villiers, and M. Ausloos,

Physica Scripta 49, 637 (1994).
[21] V. Gridin, P. Pernambuco-Wise, C.G. TrendaU, W.R.

Datars, and J.D. Garrett, Phys. Rev. B 40, 8814 (1989).
[22] H.-C. Ri, F. Kober, R. Gross, R.P. Huebener, and A.

Gupta, Phys. Rev. B 43, 13739 (1991).
[23] M. Ausloos and Ch. Laurent, Phys. Rev. B 37, 611

(1988).
[24] L. Reggiani, R. Vaglio, and A. A. Varlamov, Phys. Rev.

B 44, 9541 (1991).

[25] W. Holm, Y. Ehsev, and O. Rapp, Phys. Rev. B 51,
11992 (1995).

[26] P. Clippe, Ch. Laurent, S.K. Patapis, and M. Ausloos,
Phys. Rev. B 42, 8611 (1990).

[27] O. Cabeza, A. Pomar, A. Diaz, C. Torron, J.A. Veira, J.
Maza, and F. Vidal, Phys. Rev. B 47, 5332 (1993).

[28] J.L. Cohn, E.F. Skehon, S.A. Wolf, J.Z. Liu, and R.N.
Shehon, Phys. Rev. B 45, 13144 (1992).

[29] M. Houssa, H. Bougrine, S. Stassen, R. Cloots, and M.
Ausloos, Phys. Rev. B 54, R6885 (1996).

[30] M. Houssa, M. Ausloos, R. Cloots, and H. Bougrine,
Phys. Rev. B 56, 802 (1997).

[31] M. Ausloos, S.K. Patapis and P. Clippe, in Physics and
Materials Science of High Temperature Superconductors
II, R. Kossowsky, B. Raveau, D. Wohlleben, and S.K.
Patapis, eds., vol. 209E in the NATO ASI Series (Kluwer,
Dordrecht, 1992) pp. 755-785.

[32] A. A. Varlamov and M. Ausloos, in Fluctuation Phenom-
ena in High Temperature Superconductors, M. Ausloos
and A. A. Varlamov, eds., vol. 32 in the NATO ASI Part-
nership Sub-Series (Kluwer, Dordrecht, 1997) pp. 3-41.

[33] H.E. Stanley, Introduction to Phase Transitions and Crit-
ical Phenomena (Clarendon Press, Oxford, 1968).